摘要：随着计算机技术和信息技术的迅速发展，新闻媒介从最原始的报纸等到如今的互联网、通讯社和手机，极大的丰富了人们的生活。新闻文本可以正确的了解舆论的走向，能够快速的挖掘民众的兴趣等。越来越多的民众参与到网络中来发表自己对某一新闻事件的观点和意见。这些主观性较强的新闻文本，特别是一些突发和敏感的国内外事件，会在很短的时间内引起大家的关注，使得这些网络舆情成为隐患。因此重要的一点就是情感分析，计算机能否快速地对这些网络舆情进行情感分析，有着重大的意义。78671

本设计主要就是判断新闻文本的情感倾向性，即对新闻文本中的主观性信息进行分析。就是通常所指的正面情感和负面情感。例如“乐观”和“愉快”是褒义词，表达的是正面情感，“沮丧”和“厌恶”是贬义词，表达的是负面情感。根据传统文本情感分析的经验，对新闻文本的情感倾向进行分析。首先是文本处理，采用台湾大学整理的情感词典作为情感倾向分析的依据，算法采用最大向前字符串匹配对新闻文本进行相应的预处理及分词；然后运用朴素贝叶斯（Naive Bayes）对新闻文本进行情感倾向分析。

毕业论文关键词：新闻文本，情感词典，情感倾向，朴素贝叶斯

Abstract:With the rapid development of computer technology and information technology, the news media from the original newspapers to radio, television, Internet, and mobile phone news agency, greatly enriched the people's life。 The news text can correctly understand the direction of public opinion, can quickly mining public interest。 More and more people to participate in the network to express their views and opinions on a certain event。 The strong subjectivity of news text, especially some unexpected and sensitive events at home and abroad, will attract attention in a very short period of time, the network public opinion has become hidden。 So the important point is whether the emotional sentiment analysis, computer analysis of these the network of public opinion quickly, is of great significance。

    Sentiment analysis also known as opinion mining, sentiment is the main judgment of news text, namely the subjective information in news texts were analyzed。 Positive emotion and negative emotion is usually referred to as "optimistic。" and "happy" is the expression of commendatory terms, positive emotion, "depressed" and "disgust" is a derogatory term, is the expression of negative emotions。 According to the analysis of the traditional text emotional experience, emotional tendency of news text analysis。 First text processing, by National Taiwan University consolidation of sentiment lexicon as sentiment analysis basis, the algorithm uses the maximum forward string matching news text corresponding preprocessing and segmentation; then use naive Bayes on news text for sentiment analysis。

Key words：news text, sentiment dictionary, emotional tendency

目  录

1 引言 4

1。1 研究背景和意义 4

1。2  文本情感分析研究现状 5

1。3  系统开发工具简介 6

1。4  开发环境简介 7

1。5  论文的主要研究内容 7

1。6 论文的组织结构 8

2 相关内容介绍与理论概述 8

2。1 新闻相关概述 8

2。2  文本预处理技术